Applied Data

Apply numerical values to agree data

# Vector of expected responses
expected_responses <- c("Strongly disagree", "Disagree", "Neither disagree nor agree", "Agree", "Strongly agree")

# Ensure expected_responses is treated as a set
expected_responses_set <- sort(unique(expected_responses))

contains_only_responses <- function(column) {
  # Filter out empty and NA values
  valid_entries <- column[column != "" & !is.na(column)]
  unique_values <- sort(unique(valid_entries))
  
  # Compare lengths first to avoid recycling
  if (length(unique_values) != length(expected_responses_set)) {
    return(FALSE)
  }
  
  # Then check if all values match
  all(unique_values == expected_responses_set)
}

for (col_name in names(applied_data)) {
  if (contains_only_responses(applied_data[[col_name]])) {
    print(paste("Processing column:", col_name))

    # Replace empty strings with NA
    applied_data[[col_name]][applied_data[[col_name]] == ""] <- NA

    # Convert to factor ensuring NA handling
    applied_data[[col_name]] <- factor(applied_data[[col_name]], levels = expected_responses, ordered = TRUE)
    applied_data[[col_name]] <- as.integer(applied_data[[col_name]])
  }
}

## [1] "Processing column: P_valuable"
## [1] "Processing column: P_habit_difficult"
## [1] "Processing column: P_procrastinate"
## [1] "Processing column: P_frustration"
## [1] "Processing column: WB_use"
## [1] "Processing column: WB_plan"
## [1] "Processing column: WB_like"
## [1] "Processing column: BP_use"
## [1] "Processing column: BP_plan"
## [1] "Processing column: BP_like"
## [1] "Processing column: CC_use"
## [1] "Processing column: CC_plan"
## [1] "Processing column: CC_like"
## [1] "Processing column: APP_use"
## [1] "Processing column: APP_plan"
## [1] "Processing column: APP_like"
## [1] "Processing column: STATS_use"
## [1] "Processing column: STATS_plan"
## [1] "Processing column: STATS_like"
## [1] "Processing column: POP_use"
## [1] "Processing column: POP_plan"
## [1] "Processing column: POP_like"
## [1] "Processing column: STAPP_use"
## [1] "Processing column: STAPP_plan"
## [1] "Processing column: STAPP_like"
## [1] "Processing column: STAPP_STAPP"

applied_data_revalued <- applied_data

special_cases_column_rename_func <- function(data, column_name){
  # Ensure column_name is character type
  column_name <- as.character(column_name)
  
  # Print initial data type and summary
  cat("Initial Data Type:", class(data[[column_name]]), "\n")
  cat("Initial Data Summary:\n")
  print(summary(data[[column_name]]))
  
  # Replace empty strings with NA. Here we need to make sure that we only change empty strings.
  data[[column_name]][data[[column_name]] == ""] <- NA
  
  # It is critical to check if there are non-NA entries left after this operation.
  if (all(is.na(data[[column_name]]))) {
    cat("Warning: All values in", column_name, "have been converted to NA!\n")
  } else {
    # Convert to factor, ensuring all levels are included and ordered
    if (!all(is.na(data[[column_name]]))) {  # Only convert if there are non-NA values
      data[[column_name]] <- factor(data[[column_name]], levels = expected_responses, ordered = TRUE)
      # Convert the ordered factors to integers
      data[[column_name]] <- as.integer(data[[column_name]])
    }
  }
  
  # Print data after conversion
  cat("Data Type After Conversion:", class(data[[column_name]]), "\n")
  cat("Data Summary After Conversion:\n")
  print(summary(data[[column_name]]))
  
  # Explicitly return the modified data
  return(data)
}

# Apply the function and capture the modified data
applied_data <- special_cases_column_rename_func(applied_data, "APP_like")

## Initial Data Type: integer 
## Initial Data Summary:
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.000   4.000   5.000   4.432   5.000   5.000 
## Data Type After Conversion: integer 
## Data Summary After Conversion:
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
##      NA      NA      NA     NaN      NA      NA      74

# Save the applied file
write.csv(applied_data, "/Users/sofiascharf-matthiesen/Downloads/Applied_revalued_renamed_data_74.csv", row.names = FALSE)

Plotting

# creating function for plots
plot_response_distribution <- function(data, column_name) {
  if(any(data[[column_name]] %in% 1:5)) { # columns with Likert-scale answers
    plot_data <- as.data.frame(table(data[[column_name]]))
    names(plot_data) <- c("Response", "Frequency")
    
    # Create the plot
    p <- ggplot(plot_data, aes(x = Response, y = Frequency, fill = Response)) +
      geom_bar(stat = "identity") +
      geom_text(aes(label = Frequency), vjust = 1.5, color = "orange", size = 6) +
      labs(title = paste("Response Distribution for", column_name),
           x = "Response",
           y = "Frequency") +
      scale_fill_brewer(palette = "Pastel2") +
      theme_minimal() +
      theme(plot.title = element_text(hjust = 0.5, face = "bold", size = 10),
            panel.background = element_rect(fill = "#75bac2"),
            plot.background = element_rect(fill = "#75bac2", color = "#75bac2"))
    
    # Save the plot
    file_name <- paste0("plot_", column_name, ".png")  # filename based on column name
    ggsave(file_name, plot = p, width = 10, height = 8, units = "cm")
    
    return(p)  # Return the plot object in case it's needed elsewhere
  } else {
    print(paste("No valid data in column:", column_name))
    return(NULL)  # Return NULL if no valid data is found
  }
}

# Apply the function to all columns in the dataframe
applied_plots <- lapply(names(applied_data_revalued), function(col) plot_response_distribution(applied_data, col))

## [1] "No valid data in column: Time"
## [1] "No valid data in column: Age"
## [1] "No valid data in column: Gender"
## [1] "No valid data in column: Occupation"
## [1] "No valid data in column: ADHD"
## [1] "No valid data in column: P_tools"
## [1] "No valid data in column: Often_p_tools"

## [1] "No valid data in column: NP_no_need"
## [1] "No valid data in column: NP_idk"

## [1] "No valid data in column: APP_like"

## [1] "No valid data in column: SUGGESTIONS"
## [1] "No valid data in column: FLAWS"

applied_plots

## [[1]]
## NULL
## 
## [[2]]
## NULL
## 
## [[3]]
## NULL
## 
## [[4]]
## NULL
## 
## [[5]]
## NULL
## 
## [[6]]
## NULL
## 
## [[7]]
## NULL
## 
## [[8]]

## 
## [[9]]

## 
## [[10]]

## 
## [[11]]

## 
## [[12]]
## NULL
## 
## [[13]]
## NULL
## 
## [[14]]

## 
## [[15]]

## 
## [[16]]

## 
## [[17]]

## 
## [[18]]

## 
## [[19]]

## 
## [[20]]

## 
## [[21]]

## 
## [[22]]

## 
## [[23]]

## 
## [[24]]

## 
## [[25]]
## NULL
## 
## [[26]]

## 
## [[27]]

## 
## [[28]]

## 
## [[29]]

## 
## [[30]]

## 
## [[31]]

## 
## [[32]]

## 
## [[33]]

## 
## [[34]]

## 
## [[35]]

## 
## [[36]]
## NULL
## 
## [[37]]
## NULL

Plotting ADHD

library(dplyr)
library(ggplot2)

# Prepare the data
ADHD_data <- applied_data %>%
  select(ADHD) %>%
  group_by(ADHD) %>%
  summarise(Count = n()) %>%
  ungroup()

# Plotting as a pie chart
ADHD_pie_chart <- ggplot(ADHD_data, aes(x = "", y = Count, fill = ADHD)) +
  geom_bar(stat = "identity", width = 1) +
  coord_polar(theta = "y") +  # Convert the bar chart to a pie chart
  geom_text(aes(label = Count), position = position_stack(vjust = 0.5)) +  # Add labels in the middle of slices
  scale_fill_brewer(palette = "Pastel2") +  # Color palette
  labs(title = "ADHD Distribution",
       x = NULL,
       y = NULL) +
  theme_void() +  # Remove most non-data ink
  theme(plot.title = element_text(hjust = 0.5, face = "bold", size = 10),
        plot.background = element_rect(fill = "#75bac2", color = "#75bac2"),  # Set plot background color
        panel.background = element_rect(fill = "#75bac2", color = "#75bac2"),  # Set panel background color
        plot.margin = unit(c(0, 0, 0, 0), "cm"))  # Remove margins around the plot


ggsave("ADHD_pie_chart.png", plot = ADHD_pie_chart, width = 10, height = 8, units = "cm")

plot_response_comparison <- function(data, column_name) {
  if (any(data[[column_name]] %in% 1:5) && "ADHD" %in% names(data)) {
    # Create a subset for each ADHD group
    data_yes <- data[data$ADHD == "Yes", ]
    data_no <- data[data$ADHD == "No", ]
    
    # Creating a combined dataframe for plotting
    plot_data_yes <- as.data.frame(table(factor(data_yes[[column_name]], levels = 1:5)))
    plot_data_no <- as.data.frame(table(factor(data_no[[column_name]], levels = 1:5)))
    
    names(plot_data_yes) <- c("Response", "Frequency")
    names(plot_data_no) <- c("Response", "Frequency")
    
    plot_data_yes$ADHD <- "Yes"
    plot_data_no$ADHD <- "No"
    
    combined_plot_data <- rbind(plot_data_yes, plot_data_no)
    
    # Create the plot
    p <- ggplot(combined_plot_data, aes(x = Response, y = Frequency, fill = ADHD)) +
      geom_bar(stat = "identity", position = position_dodge(width = 0.9)) +
      geom_text(aes(label = Frequency), position = position_dodge(width = 0.9), vjust = -0.5, color = "black", size = 3.5) +
      scale_x_discrete(limits = c("1", "2", "3", "4", "5")) +  # Ensure the order is correct
      labs(title = paste("Response Distribution for", column_name, "by ADHD Status"),
           x = "Response",
           y = "Frequency") +
      scale_fill_brewer(palette = "Pastel2") +
      theme_minimal() +
      theme(plot.title = element_text(hjust = 0.5, face = "bold", size = 12),
            panel.background = element_rect(fill = "#d0e0e3"),
            plot.background = element_rect(fill = "#d0e0e3", color = "#d0e0e3"))
    
    # Save the plot
    file_name <- paste0("plot_comparison_", column_name, ".png")  # filename based on column name
    ggsave(file_name, plot = p, width = 12, height = 10, units = "cm")
    
    return(p)  # Return the plot object in case it's needed elsewhere
  } else {
    print(paste("No valid data or missing 'ADHD' column in:", column_name))
    return(NULL)  # Return NULL if no valid data is found
  }
}

# Apply the function to all columns in the dataframe
comparison_plots <- lapply(names(applied_data_revalued), function(col) plot_response_comparison(applied_data_revalued, col))

## [1] "No valid data or missing 'ADHD' column in: Time"
## [1] "No valid data or missing 'ADHD' column in: Age"
## [1] "No valid data or missing 'ADHD' column in: Gender"
## [1] "No valid data or missing 'ADHD' column in: Occupation"
## [1] "No valid data or missing 'ADHD' column in: ADHD"
## [1] "No valid data or missing 'ADHD' column in: P_tools"
## [1] "No valid data or missing 'ADHD' column in: Often_p_tools"

## [1] "No valid data or missing 'ADHD' column in: NP_no_need"
## [1] "No valid data or missing 'ADHD' column in: NP_idk"

## [1] "No valid data or missing 'ADHD' column in: SUGGESTIONS"
## [1] "No valid data or missing 'ADHD' column in: FLAWS"

# Load necessary libraries
library(ggplot2)
library(zip)

## 
## Attaching package: 'zip'

## The following objects are masked from 'package:utils':
## 
##     unzip, zip

# Define the prefix and suffix mappings
prefix_mapping <- list(
  WB = "Whiteboard",
  BP = "Block-planning",
  CC = "Colour-coding",
  APP = "App",
  STATS = "Statistics page",
  POP = "Pop-up messages",
  STAPP = "The ability to use the app without the whiteboard",
  P = "Planning"
)

suffix_mapping <- list(
  use = "I would use this feature",
  plan = "This feature would help me in my planning",
  like = "This feature would help my overall liking of the product",
  STAPP = "I would use the app as a standalone product",
  valuable = "I find it valuable",
  habit_difficult = "I find it difficult creating a habit of doing it",
  procrastinate = "I sometimes procrastinate doing it",
  frustration = "I sometimes get frustrated while doing it"
)

# Function to create descriptive column names
get_descriptive_name <- function(column_name) {
  prefix <- sub("_.*", "", column_name)
  suffix <- sub(".*_", "", column_name)
  
  prefix_desc <- prefix_mapping[[prefix]]
  suffix_desc <- suffix_mapping[[suffix]]
  
  if (!is.null(prefix_desc) && !is.null(suffix_desc)) {
    return(paste(prefix_desc, "-", suffix_desc))
  } else {
    return(column_name)
  }
}

# Function to insert a newline character in the title if it's too long
break_title <- function(title, max_length = 50) {
  if (nchar(title) > max_length) {
    words <- unlist(strsplit(title, " "))
    half_length <- ceiling(length(words) / 2)
    return(paste(paste(words[1:half_length], collapse = " "), "\n", paste(words[(half_length+1):length(words)], collapse = " "), sep = ""))
  }
  return(title)
}

# Create a directory to save plots
dir.create("plots", showWarnings = FALSE)

# Function to plot response comparison with percentages and include mean response annotations
plot_response_comparison_percentage_mean <- function(data, column_name) {
  # Check if the column contains valid data and if 'ADHD' column exists
  if (any(data[[column_name]] %in% 1:5) && "ADHD" %in% names(data)) {
    # Create subsets of data for each group based on their ADHD status
    data_yes <- data[data$ADHD == "Yes", ]
    data_no <- data[data$ADHD == "No", ]
    
    # Create frequency tables for each group and convert to dataframe
    plot_data_yes <- as.data.frame(table(factor(data_yes[[column_name]], levels = 1:5)))
    plot_data_no <- as.data.frame(table(factor(data_no[[column_name]], levels = 1:5)))
    
    # Rename the columns for clarity
    names(plot_data_yes) <- c("Response", "Count")
    names(plot_data_no) <- c("Response", "Count")
    
    # Calculate the percentage of each response category
    plot_data_yes$Frequency <- (plot_data_yes$Count / sum(plot_data_yes$Count)) * 100
    plot_data_no$Frequency <- (plot_data_no$Count / sum(plot_data_no$Count)) * 100
    
    # Calculate the mean response for each ADHD group
    mean_yes <- mean(as.numeric(as.character(data_yes[[column_name]])), na.rm = TRUE)
    mean_no <- mean(as.numeric(as.character(data_no[[column_name]])), na.rm = TRUE)
    
    # Label the data for identification in the plot
    plot_data_yes$ADHD <- "Yes"
    plot_data_no$ADHD <- "No"
    
    # Combine the datasets for plotting
    combined_plot_data <- rbind(plot_data_yes, plot_data_no)
    
    # Get the descriptive name for the column
    descriptive_name <- get_descriptive_name(column_name)
    
    # Break the title if it's too long
    plot_title <- break_title(paste("Response Distribution for", descriptive_name, "by ADHD Status (%)"))
    
    # Create the bar plot using ggplot2
    p <- ggplot(combined_plot_data, aes(x = Response, y = Frequency, fill = ADHD)) +
      geom_bar(stat = "identity", position = position_dodge(width = 0.9)) +  # Use bar geometry with identity statistic and dodge position for clarity
      geom_text(aes(label = sprintf("%.1f%%", Frequency)), position = position_dodge(width = 0.9), vjust = -0.5, color = "black", size = 3.5) +  # Add text annotations with percentage
      scale_x_discrete(limits = c("1", "2", "3", "4", "5")) +  # Explicitly set the order of x-axis categories
      labs(title = plot_title,  # Use the modified title
           subtitle = paste("Mean Response: ADHD Yes =", sprintf("%.2f", mean_yes), ", ADHD No =", sprintf("%.2f", mean_no)),
           x = "Response",
           y = "Percentage") +
      scale_fill_brewer(palette = "Pastel2") +  # Set color palette
      theme_minimal() +  # Use minimal theme for clarity
      theme(
        plot.title = element_text(hjust = 0.5, face = "bold", size = 12.5),  # Customize plot appearance
        plot.subtitle = element_text(hjust = 0.5, size = 12),
        axis.text.x = element_text(angle = 45, hjust = 1),  # Rotate x-axis labels for better readability
        plot.margin = margin(20, 20, 20, 20)  # Add margin around the plot
      )
    
    # Save the plot to a file in the "plots" directory
    file_name <- paste0("plots/plot_percentage_comparison_mean_", column_name, ".png")  # Generate dynamic filename based on column name
    ggsave(file_name, plot = p, width = 20, height = 15, units = "cm")
    
    # Return the plot object for further use or display
    return(p)
  } else {
    # Print a message if the necessary data or column is missing
    print(paste("No valid data or missing 'ADHD' column in:", column_name))
    return(NULL)  # Return NULL if conditions are not met
  }
}

# Apply the function to all columns in the dataframe to generate plots
mean_percentage_comparison_plots <- lapply(names(applied_data_revalued), function(col) plot_response_comparison_percentage_mean(applied_data_revalued, col))

## [1] "No valid data or missing 'ADHD' column in: Time"
## [1] "No valid data or missing 'ADHD' column in: Age"
## [1] "No valid data or missing 'ADHD' column in: Gender"
## [1] "No valid data or missing 'ADHD' column in: Occupation"
## [1] "No valid data or missing 'ADHD' column in: ADHD"
## [1] "No valid data or missing 'ADHD' column in: P_tools"
## [1] "No valid data or missing 'ADHD' column in: Often_p_tools"

## [1] "No valid data or missing 'ADHD' column in: NP_no_need"
## [1] "No valid data or missing 'ADHD' column in: NP_idk"

## [1] "No valid data or missing 'ADHD' column in: SUGGESTIONS"
## [1] "No valid data or missing 'ADHD' column in: FLAWS"

# Create a zip file containing all the plots
zip::zip("plots.zip", files = list.files("plots", full.names = TRUE))

# Open the directory containing the zip file (works in RStudio on Windows)
browseURL("plots.zip")

mean_percentage_comparison_plots

## [[1]]
## NULL
## 
## [[2]]
## NULL
## 
## [[3]]
## NULL
## 
## [[4]]
## NULL
## 
## [[5]]
## NULL
## 
## [[6]]
## NULL
## 
## [[7]]
## NULL
## 
## [[8]]

## 
## [[9]]

## 
## [[10]]

## 
## [[11]]

## 
## [[12]]
## NULL
## 
## [[13]]
## NULL
## 
## [[14]]

## 
## [[15]]

## 
## [[16]]

## 
## [[17]]

## 
## [[18]]

## 
## [[19]]

## 
## [[20]]

## 
## [[21]]

## 
## [[22]]

## 
## [[23]]

## 
## [[24]]

## 
## [[25]]

## 
## [[26]]

## 
## [[27]]

## 
## [[28]]

## 
## [[29]]

## 
## [[30]]

## 
## [[31]]

## 
## [[32]]

## 
## [[33]]

## 
## [[34]]

## 
## [[35]]

## 
## [[36]]
## NULL
## 
## [[37]]
## NULL

# Function to perform Shapiro-Wilk tests for normality on response data
perform_shapiro_tests <- function(data) {
  # Prepare a list to store test results
  shapiro_results <- list()

  # Iterate over each column in the dataframe
  for (column_name in names(data)) {
    # Check if the column contains numeric data suitable for normality testing
    if (any(data[[column_name]] %in% 1:5)) {
      # Subset data for each ADHD group
      group_yes <- data[data$ADHD == "Yes", column_name, drop = FALSE]
      group_no <- data[data$ADHD == "No", column_name, drop = FALSE]
      
      # Convert the subset data to numeric, omitting NA values
      numeric_yes = as.numeric(as.character(na.omit(group_yes[[column_name]])))
      numeric_no = as.numeric(as.character(na.omit(group_no[[column_name]])))

      # Check for NAs and only proceed if there are no NAs
      if (!any(is.na(numeric_yes)) && !any(is.na(numeric_no))) {
        # Initialize placeholders for results
        test_result_yes <- NULL
        test_result_no <- NULL
        
        # Check sample sizes and perform the Shapiro-Wilk test if the size is within allowable limits
        if (length(numeric_yes) >= 3 && length(numeric_yes) <= 5000) {
          test_result_yes <- shapiro.test(numeric_yes)
        } else {
          test_result_yes <- list(statistic = NA, p.value = NA, message = "Sample size out of bounds")
        }
        
        if (length(numeric_no) >= 3 && length(numeric_no) <= 5000) {
          test_result_no <- shapiro.test(numeric_no)
        } else {
          test_result_no <- list(statistic = NA, p.value = NA, message = "Sample size out of bounds")
        }

        # Store the results using the column name as the key
        shapiro_results[[column_name]] <- list(ADHD = test_result_yes, Control = test_result_no)
      }
    }
  }

  # Return the list of Shapiro-Wilk test results
  return(shapiro_results)
}

# Apply the function to the dataframe
shapiro_results <- perform_shapiro_tests(applied_data)

# Function to perform Mann-Whitney U tests on response data between ADHD groups
perform_mann_whitney_tests <- function(data) {
  # Prepare a list to store test results
  mann_whitney_results <- list()
  
  # Iterate over each column in the dataframe
  for (column_name in names(data)) {
    # Check if the column contains numeric data suitable for the Mann-Whitney U test
    if (any(data[[column_name]] %in% 1:5)) {
      # Subset data for each ADHD group
      group_yes <- data[data$ADHD == "Yes", column_name, drop = FALSE]
      group_no <- data[data$ADHD == "No", column_name, drop = FALSE]
      
      # Perform the Mann-Whitney U test comparing the two groups
      # na.omit is used to remove NA values that can affect the test
      test_result <- wilcox.test(as.numeric(as.character(na.omit(group_yes[[1]]))), 
                                 as.numeric(as.character(na.omit(group_no[[1]]))),
                                 alternative = "two.sided",  # You can change this to "less" or "greater" if needed
                                 exact = FALSE)  # Use approximation rather than exact calculation if data size is large
      
      # Store the result using the column name as the key
      mann_whitney_results[[column_name]] <- test_result
    }
  }
  
  # Return the list of Mann-Whitney U test results
  return(mann_whitney_results)
}

# Apply the function to the applied_data dataframe
mann_whitney_results <- perform_mann_whitney_tests(applied_data_revalued)

# To view the results, you can print them or access specific results like this:
# print(mann_whitney_results[["SomeColumnName"]])  # Replace SomeColumnName with an actual column name

mann_whitney_results

## $P_valuable
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 341.5, p-value = 0.08794
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $P_habit_difficult
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 708.5, p-value = 0.0005858
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $P_procrastinate
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 622, p-value = 0.02531
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $P_frustration
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 617.5, p-value = 0.028
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $WB_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 571.5, p-value = 0.5363
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $WB_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 474, p-value = 0.5122
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $WB_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 405.5, p-value = 0.1268
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $BP_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 435.5, p-value = 0.2679
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $BP_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 502.5, p-value = 0.7887
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $BP_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 440, p-value = 0.2866
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $CC_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 529.5, p-value = 0.9322
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $CC_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 507.5, p-value = 0.8475
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $CC_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 498.5, p-value = 0.7557
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $APP_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 454, p-value = 0.3399
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $APP_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 416.5, p-value = 0.1453
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $APP_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 432, p-value = 0.2024
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STATS_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 455, p-value = 0.3872
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STATS_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 462.5, p-value = 0.4503
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STATS_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 517, p-value = 0.9488
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $POP_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 427.5, p-value = 0.2313
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $POP_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 518, p-value = 0.9595
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $POP_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 492, p-value = 0.7029
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STAPP_use
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 365, p-value = 0.03083
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STAPP_plan
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 406, p-value = 0.1152
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STAPP_like
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 432.5, p-value = 0.2288
## alternative hypothesis: true location shift is not equal to 0
## 
## 
## $STAPP_STAPP
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  as.numeric(as.character(na.omit(group_yes[[1]]))) and as.numeric(as.character(na.omit(group_no[[1]])))
## W = 467.5, p-value = 0.4803
## alternative hypothesis: true location shift is not equal to 0

library(dplyr)
library(ggplot2)
library(gridExtra)

## 
## Attaching package: 'gridExtra'

## The following object is masked from 'package:dplyr':
## 
##     combine

# Define the prefix and suffix mappings
prefix_mapping <- list(
  WB = "Whiteboard",
  BP = "Block-planning",
  CC = "Colour-coding",
  APP = "App",
  STATS = "Statistics page",
  POP = "Pop-up MSGs",
  STAPP = "Standalone App"
)

suffix_mapping <- list(
  use = "I would use this feature",
  plan = "This feature would help me in my planning",
  like = "This feature would help my overall liking of the product"
)

# Function to create descriptive column names
get_prefix_name <- function(column_name) {
  prefix <- sub("_.*", "", column_name)
  prefix_desc <- prefix_mapping[[prefix]]
  
  if (!is.null(prefix_desc)) {
    return(prefix_desc)
  } else {
    return(column_name)
  }
}

# Function to translate the suffix to a descriptive title
get_suffix_description <- function(suffix) {
  if (suffix %in% names(suffix_mapping)) {
    return(suffix_mapping[[suffix]])
  } else {
    return(suffix)
  }
}

# Function to plot mean responses for specified suffixes in column names
plot_mean_responses <- function(data, suffix_list) {
  # Loop through each suffix to create plots
  for (suffix in suffix_list) {
    # Filter columns that end with the current suffix
    relevant_columns <- names(data)[grepl(paste0(suffix, "$"), names(data))]
    
    # Prepare data frames to store mean responses for each group
    adhd_mean_responses <- data.frame(Column = character(), Mean = numeric())
    control_mean_responses <- data.frame(Column = character(), Mean = numeric())
    
    # Calculate mean responses for each relevant column
    for (column_name in relevant_columns) {
      # Extract data for each group
      adhd_data <- data[data$ADHD == "Yes", column_name, drop = FALSE]
      control_data <- data[data$ADHD == "No", column_name, drop = FALSE]
      
      # Calculate means
      adhd_mean <- mean(as.numeric(as.character(na.omit(adhd_data[[1]]))), na.rm = TRUE)
      control_mean <- mean(as.numeric(as.character(na.omit(control_data[[1]]))), na.rm = TRUE)
      
      # Append to the data frames
      adhd_mean_responses <- rbind(adhd_mean_responses, data.frame(Column = column_name, Mean = adhd_mean))
      control_mean_responses <- rbind(control_mean_responses, data.frame(Column = column_name, Mean = control_mean))
    }
    
    # Translate column names for the y-axis labels
    adhd_mean_responses$Column <- sapply(adhd_mean_responses$Column, get_prefix_name)
    control_mean_responses$Column <- sapply(control_mean_responses$Column, get_prefix_name)
    
    # Add ranking numbers to y-axis labels
    adhd_mean_responses <- adhd_mean_responses %>%
      arrange(desc(Mean)) %>%
      mutate(Rank = row_number(), Column = paste(Rank, "=", Column))
    
    control_mean_responses <- control_mean_responses %>%
      arrange(desc(Mean)) %>%
      mutate(Rank = row_number(), Column = paste(Rank, "=", Column))
    
    # Reorder the factor levels based on Rank in descending order
    adhd_mean_responses$Column <- factor(adhd_mean_responses$Column, levels = rev(adhd_mean_responses$Column))
    control_mean_responses$Column <- factor(control_mean_responses$Column, levels = rev(control_mean_responses$Column))
    
    # Translate the suffix
    suffix_desc <- get_suffix_description(sub("_", "", suffix))
    
    # Create a horizontal bar plot for ADHD group
    p_adhd <- ggplot(adhd_mean_responses, aes(x = Mean, y = Column)) +
      geom_bar(stat = "identity", fill = "#FDCDAC") +
      geom_text(aes(label = sprintf("%.1f", Mean)), hjust = -0.3, size = 3) +  # Add mean value labels
      labs(title = paste("Mean Response for ADHD Group"),
           x = "",
           y = suffix_desc, 
           size = 7) +
      xlim(0, 5) +  # Set x-axis limits
      theme_minimal() +
      theme(plot.title = element_text(hjust = 0.5, face = "bold", size = 9),
            axis.text.x = element_text(size = 8),
            plot.margin = margin(t = 20, r = 30, b = 20, l = 20))
    
    # Create a horizontal bar plot for Control group
    p_control <- ggplot(control_mean_responses, aes(x = Mean, y = Column)) +
      geom_bar(stat = "identity", fill = "#B3E2CD") +
      geom_text(aes(label = sprintf("%.1f", Mean)), hjust = -0.3, size = 3) +  # Add mean value labels
      labs(title = paste("Mean Response for Control Group"),
           x = "",
           y = suffix_desc, 
           size = 7) +
      xlim(0, 5) +  # Set x-axis limits
      theme_minimal() +
      theme(plot.title = element_text(hjust = 0.5, face = "bold", size = 9),
            axis.text.x = element_text(size = 8),
            plot.margin = margin(t = 20, r = 30, b = 20, l = 20))
    
    # Arrange the two plots in one window using gridExtra with wider dimensions and smaller heights
    combined_plot <- grid.arrange(p_adhd, p_control, ncol = 2, widths = c(3, 3), heights = c(0.5))
    
    # Save the combined plot
    ggsave(paste0("mean_response_", suffix, ".png"), plot = combined_plot, width = 14, height = 7, dpi = 300)
  }
}

# Apply the function with the specified suffixes
plot_mean_responses(applied_data_revalued, c("_use", "_plan", "_like"))